@jcheesepkg/nanobot 0.9.0 → 0.9.2

package/skills/cron/SKILL.md

@@ -15,21 +15,25 @@ Use the `cron` tool to schedule reminders or recurring tasks.
 ## Examples
 
 Fixed reminder (kind="direct", the default):
+
 ```
 cron(action="add", message="Time to take a break!", every_seconds=1200)
 ```
 
 Dynamic task (kind="task", agent executes each time):
+
 ```
 cron(action="add", message="Check HKUDS/nanobot GitHub stars and report", kind="task", every_seconds=600)
 ```
 
 Daily at 8am (user is in Asia/Tokyo):
+
 ```
 cron(action="add", message="おはよう！今日も頑張ろう", cron_expr="0 8 * * *", timezone="Asia/Tokyo")
 ```
 
 List/remove:
+
 ```
 cron(action="list")
 cron(action="remove", job_id="abc123")
@@ -49,11 +53,11 @@ For `every_seconds`, no timezone is needed — it's just an interval.
 
 ## Time Expressions
 
-| User says | Parameters |
-|-----------|------------|
-| every 20 minutes | every_seconds: 1200 |
-| every hour | every_seconds: 3600 |
-| every day at 8am | cron_expr: "0 8 * * *", timezone: (from system prompt) |
-| weekdays at 5pm | cron_expr: "0 17 * * 1-5", timezone: (from system prompt) |
-| In 10 minutes | at: (current time + 10min, with tz offset) |
-| at a specific date/time | at: ISO-8601 with timezone offset |
+| User says               | Parameters                                                  |
+| ----------------------- | ----------------------------------------------------------- |
+| every 20 minutes        | every_seconds: 1200                                         |
+| every hour              | every_seconds: 3600                                         |
+| every day at 8am        | cron_expr: "0 8 \* \* \*", timezone: (from system prompt)   |
+| weekdays at 5pm         | cron_expr: "0 17 \* \* 1-5", timezone: (from system prompt) |
+| In 10 minutes           | at: (current time + 10min, with tz offset)                  |
+| at a specific date/time | at: ISO-8601 with timezone offset                           |

package/skills/daily-summary/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: daily-summary
-description: Morning briefing with weather and schedule.
+description: Automated morning briefing combining weather, schedule, and trash day. Use when user asks for 朝の要約 or wants a daily digest setup.
 ---
 
 # Daily Summary (朝の要約)
@@ -31,6 +31,7 @@ curl -s "wttr.in/Tokyo?format=j1" | jq -r '
 ### 2. Schedule
 
 Check for events:
+
 ```bash
 grep "$(date +%Y-%m-%d)" memory/HISTORY.md 2>/dev/null
 ```
@@ -38,6 +39,7 @@ grep "$(date +%Y-%m-%d)" memory/HISTORY.md 2>/dev/null
 ### 3. Trash Day
 
 Day of week:
+
 - 火・金 = 燃えるゴミ
 - 水 = 燃えないゴミ
 - 土 = ペットボトル (varies by ward)
@@ -47,6 +49,7 @@ Day of week:
 **ALWAYS use the `flex_message` tool** to display the daily summary as a rich card.
 
 Call the tool like this:
+
 ```
 flex_message(template="morning_summary", data={
   greeting: "おはよう！",
@@ -63,6 +66,7 @@ The card is sent automatically — do NOT repeat the JSON. Just respond naturall
 ## Personalization
 
 Read from `memory/MEMORY.md`:
+
 - `city:` for weather location
 - `ward:` for trash schedule
 - `wake_time:` for optimal push time

package/skills/english/SKILL.md

@@ -1,18 +1,28 @@
 ---
 name: english
-description: Casual English conversation practice with corrections.
+description: English conversation practice with level-based corrections, vocab tracking, and quizzes. Use when user says 英語, English mode, 英会話, 英語クイズ, or vocab test.
 ---
 
 # English Chat (英会話練習)
 
+## Data
+
+All data in `data/english/`:
+
+- `settings.json` — level preference
+- `vocab.json` — accumulated vocabulary list
+- `sessions.jsonl` — session logs (one JSON object per line)
+
 ## Difficulty Level
 
-Settings stored in `data/english.json`:
+Settings stored in `data/english/settings.json`:
+
 ```json
 { "level": "beginner" }
 ```
 
-On first activation, if `data/english.json` doesn't exist, ask the user to choose their level:
+On first activation, if `data/english/settings.json` doesn't exist, ask the user to choose their level:
+
 ```
 flex_message(template="action_buttons", data={
   prompt: "英語レベルを選んでね！\nChoose your English level:",
@@ -24,22 +34,22 @@ flex_message(template="action_buttons", data={
 })
 ```
 
-Save the choice to `data/english.json`. User can change anytime with "レベル変更".
+Save the choice to `data/english/settings.json`. User can change anytime with "レベル変更".
 
 ### Level behavior
 
-| | Beginner | Intermediate | Advanced |
-|---|---|---|---|
-| Bot's vocab | Simple, common words | Natural, varied | Idioms, slang, nuance |
-| Corrections | Only major grammar | Grammar + word choice | Everything incl. nuance |
-| Japanese help | Always add 日本語 hints | On request only | Never unless asked |
-| Sentence length | Short, simple | Natural | Complex, compound |
+|                 | Beginner                | Intermediate          | Advanced                |
+| --------------- | ----------------------- | --------------------- | ----------------------- |
+| Bot's vocab     | Simple, common words    | Natural, varied       | Idioms, slang, nuance   |
+| Corrections     | Only major grammar      | Grammar + word choice | Everything incl. nuance |
+| Japanese help   | Always add 日本語 hints | On request only       | Never unless asked      |
+| Sentence length | Short, simple           | Natural               | Complex, compound       |
 
 ## Mode
 
 When user says "英語モード" / "English mode" / "英会話":
 
-1. Read `data/english.json` for level (default: beginner if missing)
+1. Read `data/english/settings.json` for level (default: beginner if missing)
 2. Reply in English, adapted to level
 3. Gently correct mistakes (strictness based on level)
 4. Offer Japanese translation based on level rules above
@@ -51,11 +61,12 @@ Casual and natural. Like chatting with a friend.
 User: "Yesterday I go to Tokyo"
 Bot: "Oh nice! Where did you go in Tokyo? 🗼
 
-(Correction: 'Yesterday I *went* to Tokyo' - past tense!)"
+(Correction: 'Yesterday I _went_ to Tokyo' - past tense!)"
 
 ## Corrections
 
 Only fix:
+
 - Major grammar (tense, word order)
 - Wrong word usage
 - Natural phrasing
@@ -65,6 +76,7 @@ Adjust strictness per level. Skip minor stuff. Keep flow.
 ## Topic Selection
 
 Use quick_replies directive:
+
 ```
 What do you want to practice?
 
@@ -74,14 +86,17 @@ What do you want to practice?
 ## Topic Prompts
 
 **Daily life:**
+
 - "What did you have for breakfast?"
 - "How do you spend weekends?"
 
 **Travel:**
+
 - "Where do you want to go next?"
 - "Best place you've visited?"
 
 **Work:**
+
 - "What do you do?"
 - "Remote or office?"
 
@@ -93,6 +108,7 @@ What do you want to practice?
 ## Code-switching
 
 If user struggles:
+
 ```
 How was your day? (今日どうだった？)
 ```
@@ -106,12 +122,30 @@ When user says "日本語に戻して" / "exit English" / "おわり":
    - Corrections you made (original → fixed)
    - New vocabulary or phrases you introduced
    - An encouraging comment about their performance
-2. Send a session summary using `flex_message`:
+
+2. **Save session log** — append one JSON line to `data/english/sessions.jsonl`:
+
+```bash
+echo '{"date":"2025-02-13","topics":["daily life","food"],"corrections":[{"wrong":"I go","right":"I went","rule":"past tense"}],"vocab":["commute","grab a bite"],"duration_msgs":12}' >> data/english/sessions.jsonl
+```
+
+3. **Update vocab list** — read `data/english/vocab.json`, merge new words, write back:
+
+```json
+[
+  { "word": "commute", "meaning": "通勤する", "date": "2025-02-13" },
+  { "word": "grab a bite", "meaning": "軽く食べる", "date": "2025-02-13" }
+]
+```
+
+If the file doesn't exist, create it with the new entries. Skip duplicates (same `word`).
+
+4. **Send session summary card**:
 
 ```
 flex_message(template="custom", data={
   title: "📝 English Session",
-  body: "Topics: ...\n\nCorrections:\n❌ I go → ✅ I went\n❌ more easy → ✅ easier\n\nNew vocab:\n• commute (通勤する)\n• grab a bite (軽く食べる)\n\n💪 Great job! Your past tense is improving!",
+  body: "Topics: daily life, food\n\nCorrections:\n❌ I go → ✅ I went\n❌ more easy → ✅ easier\n\nNew vocab:\n• commute (通勤する)\n• grab a bite (軽く食べる)\n\n💪 Great job! Your past tense is improving!",
   header_color: "#4CAF50"
 })
 ```
@@ -120,6 +154,30 @@ Keep it honest — if there were no corrections, say so. If there were many, pic
 
 After the card, respond: "英語練習お疲れ様！👍"
 
+## Vocab Review
+
+When user says "英語クイズ" / "vocab test" / "復習":
+
+1. Read `data/english/vocab.json`
+2. Pick 5 random words from past sessions
+3. Quiz the user — show the English word, ask for meaning or usage in a sentence
+4. Use `action_buttons` for multiple choice if beginner level
+
+Example:
+
+```
+flex_message(template="action_buttons", data={
+  prompt: "What does 'commute' mean?",
+  buttons: [
+    { label: "通勤する", text: "答え 通勤する", style: "primary" },
+    { label: "通信する", text: "答え 通信する", style: "secondary" },
+    { label: "通過する", text: "答え 通過する", style: "secondary" }
+  ]
+})
+```
+
+For intermediate/advanced: just ask in text, no multiple choice.
+
 ## Tips
 
 - Match user's level

package/skills/expense/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: expense
-description: Track daily expenses and show monthly summaries.
+description: Expense tracker with auto-categorization and monthly receipt summaries. Use when user mentions spending like 出費, 500円, or asks for expense reports.
 ---
 
 # Expense Tracker (出費記録)
@@ -10,21 +10,22 @@ description: Track daily expenses and show monthly summaries.
 User says: "コンビニで500円" / "出費 +380 カフェ" / "ランチ 1200"
 
 Log to CSV:
+
 ```bash
-echo "$(date +%Y-%m-%d),$(date +%H:%M),500,コンビニ,コンビニ" >> data/expenses.csv
+echo "$(date +%Y-%m-%d),$(date +%H:%M),500,コンビニ,コンビニ" >> data/expense/log.csv
 ```
 
 Format: `date,time,amount,note,category`
 
 ## Auto-categorize
 
-| Keywords | Category |
-|----------|----------|
-| コンビニ, セブン, ローソン, ファミマ | コンビニ |
-| カフェ, スタバ, ドトール | カフェ |
-| ランチ, 食事, 丼, ラーメン | 食事 |
-| 電車, 地下鉄, バス | 交通 |
-| amazon, ヨドバシ | ショッピング |
+| Keywords                             | Category     |
+| ------------------------------------ | ------------ |
+| コンビニ, セブン, ローソン, ファミマ | コンビニ     |
+| カフェ, スタバ, ドトール             | カフェ       |
+| ランチ, 食事, 丼, ラーメン           | 食事         |
+| 電車, 地下鉄, バス                   | 交通         |
+| amazon, ヨドバシ                     | ショッピング |
 
 Default: 雑費
 
@@ -35,18 +36,21 @@ After recording, confirm with a short text: "✅ 記録: カフェ ¥380"
 ## Summary
 
 Monthly totals:
+
 ```bash
-awk -F, '$1 ~ /^2025-02/ {sum+=$3} END {print sum}' data/expenses.csv
+awk -F, '$1 ~ /^2025-02/ {sum+=$3} END {print sum}' data/expense/log.csv
 ```
 
 By category:
+
 ```bash
-awk -F, '$1 ~ /^2025-02/ {cat[$5]+=$3} END {for(c in cat) print c, cat[c]}' data/expenses.csv
+awk -F, '$1 ~ /^2025-02/ {cat[$5]+=$3} END {for(c in cat) print c, cat[c]}' data/expense/log.csv
 ```
 
 **ALWAYS use `flex_message`** for summaries. The card is auto-sent.
 
 Monthly summary as receipt card:
+
 ```
 flex_message(template="receipt", data={
   title: "2月の出費",

package/skills/fortune/SKILL.md

@@ -1,30 +1,31 @@
 ---
 name: fortune
-description: Daily horoscope readings.
+description: Daily horoscope and zodiac readings. Use when user says 占い, 運勢, 星座, or asks about their fortune/horoscope.
 ---
 
 # Fortune (占い)
 
 ## 12 Signs
 
-| Sign | Dates | Emoji |
-|------|-------|-------|
-| 牡羊座 | 3/21-4/19 | ♈ |
-| 牡牛座 | 4/20-5/20 | ♉ |
-| 双子座 | 5/21-6/21 | ♊ |
-| 蟹座 | 6/22-7/22 | ♋ |
-| 獅子座 | 7/23-8/22 | ♌ |
-| 乙女座 | 8/23-9/22 | ♍ |
-| 天秤座 | 9/23-10/23 | ♎ |
-| 蠍座 | 10/24-11/22 | ♏ |
-| 射手座 | 11/23-12/21 | ♐ |
-| 山羊座 | 12/22-1/19 | ♑ |
-| 水瓶座 | 1/20-2/18 | ♒ |
-| 魚座 | 2/19-3/20 | ♓ |
+| Sign   | Dates       | Emoji |
+| ------ | ----------- | ----- |
+| 牡羊座 | 3/21-4/19   | ♈    |
+| 牡牛座 | 4/20-5/20   | ♉    |
+| 双子座 | 5/21-6/21   | ♊    |
+| 蟹座   | 6/22-7/22   | ♋    |
+| 獅子座 | 7/23-8/22   | ♌    |
+| 乙女座 | 8/23-9/22   | ♍    |
+| 天秤座 | 9/23-10/23  | ♎    |
+| 蠍座   | 10/24-11/22 | ♏    |
+| 射手座 | 11/23-12/21 | ♐    |
+| 山羊座 | 12/22-1/19  | ♑    |
+| 水瓶座 | 1/20-2/18   | ♒    |
+| 魚座   | 2/19-3/20   | ♓    |
 
 ## Generate Fortune
 
 Random daily seed:
+
 ```bash
 seed=$(date +%Y%m%d)$sign_index
 fortune_level=$((seed % 5 + 1))  # 1-5 stars
@@ -32,19 +33,20 @@ fortune_level=$((seed % 5 + 1))  # 1-5 stars
 
 Fortune messages:
 
-| Stars | Message |
-|-------|---------|
-| ★★★★★ | 最高の日！自信を持って |
+| Stars | Message                      |
+| ----- | ---------------------------- |
+| ★★★★★ | 最高の日！自信を持って       |
 | ★★★★☆ | いい流れ。チャンスを掴んで！ |
-| ★★★☆☆ | 普通。マイペースで |
-| ★★☆☆☆ | 注意深く。深呼吸して |
-| ★☆☆☆☆ | 慎重に。無理は禁物 |
+| ★★★☆☆ | 普通。マイペースで           |
+| ★★☆☆☆ | 注意深く。深呼吸して         |
+| ★☆☆☆☆ | 慎重に。無理は禁物           |
 
 ## Output
 
 **ALWAYS use the `flex_message` tool** to display fortune results as a rich card.
 
 Call the tool like this:
+
 ```
 flex_message(template="fortune", data={
   sign: "♍ 乙女座",
@@ -65,6 +67,7 @@ The card is sent automatically — do NOT repeat the JSON. Just respond naturall
 ## Personalization
 
 Store in `memory/MEMORY.md`:
+
 ```
 - 星座: 牡羊座
 - 誕生日: 4月15日
@@ -79,6 +82,7 @@ cron(action="add", message="今日の運勢見る？", cron_expr="0 8 * * *", ti
 ## Personality
 
 Match bot tone:
+
 - フレンドリー: message="今日はラッキーかも！"
 - 執事: message="本日の運勢でございます"
 - ツンデレ: message="べ、別に...星4つ"

package/skills/habit/SKILL.md

@@ -1,13 +1,14 @@
 ---
 name: habit
-description: Track daily habits with streaks.
+description: Habit tracker with streak counting and weekly summaries. Use when user says 習慣, mentions completing a habit, or wants to add/check habits.
 ---
 
 # Habit Tracker (習慣トラッカー)
 
 ## Setup
 
-Store habits in `data/habits.json`:
+Store habits in `data/habit/config.json`:
+
 ```json
 {
   "habits": [
@@ -45,19 +46,19 @@ The card is auto-sent — do NOT repeat the JSON.
 
 - Green header (#4CAF50) = already done today
 - Orange header (#FF9800) = not done yet
-- Read `data/habit-log.csv` to check today's status before building the carousel
+- Read `data/habit/log.csv` to check today's status before building the carousel
 
 ## Logging
 
 ```bash
-echo "$(date +%Y-%m-%d)|meditate|1" >> data/habit-log.csv
+echo "$(date +%Y-%m-%d)|meditate|1" >> data/habit/log.csv
 ```
 
 ## Streaks
 
 ```bash
 # Count consecutive days
-awk -F"|" '$2=="meditate" && $3=="1" {print $1}' data/habit-log.csv | uniq | tail -7 | wc -l
+awk -F"|" '$2=="meditate" && $3=="1" {print $1}' data/habit/log.csv | uniq | tail -7 | wc -l
 ```
 
 ## Summary
@@ -82,4 +83,4 @@ flex_message(template="custom", data={
 ## Motivation
 
 - Streak celebration: "🔥 7日連続すごい！"
-- Gentle nudge: "今日まだ瞑想してないよ〜"
+- Gentle nudge: "今日まだ瞑想してないよ〜"

package/skills/hydration/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: hydration
-description: Track daily water intake with reminders.
+description: Water intake tracker with reminders and progress display. Use when user says 水飲んだ, 一杯, or asks about hydration.
 ---
 
 # Hydration Tracker (水分摂取)
@@ -10,14 +10,15 @@ description: Track daily water intake with reminders.
 User says: "水飲んだ" / "一杯飲んだ" / "+1"
 
 Log:
+
 ```bash
-echo "$(date -Iseconds)|1" >> data/hydration.log
+echo "$(date -Iseconds)|1" >> data/hydration/log.csv
 ```
 
 ## Check Progress
 
 ```bash
-grep "$(date +%Y-%m-%d)" data/hydration.log | wc -l
+grep "$(date +%Y-%m-%d)" data/hydration/log.csv | wc -l
 ```
 
 ## Output
@@ -25,6 +26,7 @@ grep "$(date +%Y-%m-%d)" data/hydration.log | wc -l
 **ALWAYS use the `flex_message` tool** to display hydration status.
 
 Call the tool like this:
+
 ```
 flex_message(template="hydration", data={
   title: "水飲んだ？",
@@ -47,11 +49,12 @@ cron(action="add", message="水飲んだ？リマインド送って", every_seco
 ## Reset
 
 Archive at midnight:
+
 ```bash
-[ -f data/hydration.log ] && mv data/hydration.log data/hydration-$(date +%Y-%m-%d).log
+[ -f data/hydration/log.csv ] && mv data/hydration/log.csv data/hydration/log-$(date +%Y-%m-%d).csv
 ```
 
 ## Notes
 
 - ALWAYS use the `flex_message` tool — the card is auto-sent
-- Update `current` count by reading from `data/hydration.log` before calling the tool
+- Update `current` count by reading from `data/hydration/log.csv` before calling the tool

package/skills/memory/SKILL.md

@@ -22,6 +22,7 @@ Use the `exec` tool to run grep. Combine patterns: `grep -iE "meeting|deadline"
 ## When to Update MEMORY.md
 
 Write important facts immediately using `edit_file` or `write_file`:
+
 - User preferences ("I prefer dark mode")
 - Project context ("The API uses OAuth2")
 - Relationships ("Alice is the project lead")

package/skills/mood/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: mood
-description: Daily mood tracking with emoji and optional notes.
+description: Mood journal with emoji tracking and weekly reports. Use when user shares feelings like 気分, だるい, 疲れた, いい感じ, or asks for mood summary.
 ---
 
 # Mood Journal (気分日記)
@@ -11,16 +11,17 @@ User says: "今日はいい感じ" / "気分いい" / "だるい" / "疲れた"
 
 Map to emoji + score:
 
-| Input patterns | Emoji | Score |
-|----------------|-------|-------|
-| いい, 最高, 😊, 🥰, 嬉しい | 😊 | 5 |
-| まあまあ, 普通, 😐 | 😐 | 3 |
-| だるい, 疲れた, 😔, 悲しい | 😔 | 2 |
-| 最悪, 😢, 辛い | 😢 | 1 |
+| Input patterns             | Emoji | Score |
+| -------------------------- | ----- | ----- |
+| いい, 最高, 😊, 🥰, 嬉しい | 😊    | 5     |
+| まあまあ, 普通, 😐         | 😐    | 3     |
+| だるい, 疲れた, 😔, 悲しい | 😔    | 2     |
+| 最悪, 😢, 辛い             | 😢    | 1     |
 
 Log to CSV:
+
 ```bash
-echo "$(date +%Y-%m-%d),$(date +%H:%M),😊,5,今日はいい感じ" >> data/mood.csv
+echo "$(date +%Y-%m-%d),$(date +%H:%M),😊,5,今日はいい感じ" >> data/mood/log.csv
 ```
 
 Format: `date,time,emoji,score,note`
@@ -28,6 +29,7 @@ Format: `date,time,emoji,score,note`
 ## Daily Check-in
 
 Use `action_buttons` flex for tap response:
+
 ```
 flex_message(template="action_buttons", data={
   prompt: "今日の気分は？",
@@ -41,6 +43,7 @@ flex_message(template="action_buttons", data={
 ```
 
 The card is auto-sent. Set up cron for nightly check-in:
+
 ```
 cron(action="add", message="気分チェックインのカードを送って", cron_expr="0 21 * * *", timezone="Asia/Tokyo", kind="task")
 ```
@@ -48,8 +51,9 @@ cron(action="add", message="気分チェックインのカードを送って", c
 ## Summary
 
 Calculate average:
+
 ```bash
-awk -F, 'NR>7 {sum+=$4; count++} END {print "avg:", sum/count}' data/mood.csv
+awk -F, 'NR>7 {sum+=$4; count++} END {print "avg:", sum/count}' data/mood/log.csv
 ```
 
 **ALWAYS use `flex_message`** for weekly summaries. The card is auto-sent.

package/skills/skill-creator/SKILL.md

@@ -36,6 +36,7 @@ skill-name/
   scripts/      (optional) - Executable code
   references/   (optional) - Documentation for context
   assets/       (optional) - Files used in output
+  data/         (optional) - User data, settings, logs
 ```
 
 #### SKILL.md
@@ -72,6 +73,24 @@ Skills use three-level loading:
 - `description`: What the skill does AND when to use it. Include trigger contexts.
   All "when to use" info goes here since the body only loads after triggering.
 
+### Data Storage
+
+**Custom skills** live in the workspace `skills/` directory. Store user data in `data/` within the skill, using the **full path from workspace root** in SKILL.md:
+
+```
+skills/my-skill/
+  SKILL.md
+  data/
+    log.csv
+    settings.json
+```
+
+```bash
+echo "..." >> skills/my-skill/data/log.csv
+```
+
+NOT `data/log.csv` — that would write to the workspace root.
+
 ### Flex Message Integration
 
 If the skill produces output that would benefit from rich visual display (cards, progress trackers, summaries, confirmations), include `flex_message` tool usage in the SKILL.md.
@@ -81,6 +100,7 @@ Available templates: `fortune`, `info_card`, `action_buttons`, `receipt`, `morni
 #### Custom template
 
 Single card with auto-built layout:
+
 ```
 flex_message(template="custom", data={
   title: "カードタイトル",
@@ -96,6 +116,7 @@ flex_message(template="custom", data={
 ```
 
 Carousel (horizontally swipeable cards):
+
 ```
 flex_message(template="custom", data={
   bubbles: [
@@ -108,6 +129,7 @@ flex_message(template="custom", data={
 #### Output section for skills using flex
 
 The card is **auto-sent** when the tool returns. Include this in SKILL.md:
+
 ```markdown
 ## Output
 

package/skills/summarize/SKILL.md

@@ -10,6 +10,7 @@ Fast CLI to summarize URLs, local files, and YouTube links.
 ## When to use
 
 Use this skill when the user:
+
 - Sends a URL and asks "what's this about?"
 - Asks to summarize an article, blog post, or web page
 - Asks to transcribe or summarize a YouTube video